A widespread SCC mec -located gene cluster protects methicillin-resistant Staphylococcus aureus against toxic polysulfides

23 staphylococci, Staphylococcus aureus , MRSA, SCCmec, cst, microbiome, ecophysiology, sulfide, 24 colonization, dysbiosis 25 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 2

26 The genus Staphylococcus contains important human commensals and pathogens, including 27 methicillin-resistant Staphylococcus aureus (MRSA), which is a frequent colonizer of humans and a 28 leading cause of healthcare -associated and life -threatening infections. While its virulence and 29 pathogenicity have been extensively studied, factors driving the colonization and distribution of MRSA 30 as a pathobiont are less understood. Here, we report on a cst sulfide detoxification gene cluster located 31 on SCC mec, the antibiotic resistance-mediating genetic element of MRSA . Bioinformatic analyses 32 revealed a heterogeneous distribution of cst clusters in staphylococcal genomes and that many clinically 33 relevant SCC mec types introduce an additional cst cluster ( cst2) to MRSA. While the canonical cst 34 cluster (cst1) consists of the five genes tauE, cstR, cstA, cstB, and sqr, most staphylococcal cst clusters, 35 including the SCCmec-located cst2, lack the sqr gene, which encodes for a sulfide:quinone reductase 36 responsible for the initial step of sulfide detoxification. Growth experiments with a diverse set of 37 representative Staphylococcus strains, cst-deletion mutants, and complementation with cst-containing 38 plasmids demonstrated that the cst cluster enables sqr-independent polysulfide -detoxification. 39 Furthermore, the additional cst2 cluster confers high polysulfide tolerance to MRSA, providing the 40 pathogen with a unique advantage in polysulfide -rich environmen ts. Using serial passaging co -41 cultivation experiments with methicillin-sensitive S. aureus (MSSA) strains, we demonstrated that in the 42 presence of polysulfides cst2-containing MRSA can invade an established MSSA population and 43 outperform the occupying resi dent in direct competition. Overall, our findings indicate that polysulfides 44 are critical stress factors for staphylococci, potentially contributing to the spread of cst2-containing 45 SCCmec and MRSA. 46 Importance 47 Methicillin-resistant Staphylococcus aureus ( MRSA) is one of the most prevalent human pathogens 48 responsible for millions of life-threatening infections worldwide. It acquires antibiotic resistance through 49 the genetic element SCCmec, which contains the characteristic mecA gene that renders the organis m 50 resistant to most classes of β-lactam antibiotics. Besides mecA and accessory gene complexes 51 necessary for the transfer of SCCmec and phenotype manifestation, the genetic element also contains 52 prominent gene clusters with unknown functions. Here, we report on a (poly-)sulfide-detoxification gene 53 cluster ( cst2) present on SCC mec that provides MRSA with a unique advantage in environments 54 containing polysulfides – highly reactive intermediates of sulfide oxidation naturally occurring as 55 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 3 microbial stressors on mucosal surfaces inside the human body. We demonstrate that in the presence 56 of polysulfides, cst2 enables MRSA to outperform non-MRSA in direct competition, thus supporting the 57 invasion and proliferation of this pathogen independent of its antibiotic resistance. 58

59 Methicillin-resistant Staphylococcus aureus (MRSA) is one of the most prevalent human pathogens, 60 responsible for millions of life-threatening infections worldwide (1). It acquires resistance to beta-lactam 61 antibiotics through the staphylococcal cassette chromosome mec (SCCmec), a mobile genetic element. 62 SCCmec contains the mec gene complex for resistance, the ccr gene complex responsible for genomic 63 integration, and variable joining regions. Different SCCmec types (I-XV) and subtypes are defined by 64 order, length, and sequence of these core compone nts. SCCmec types I-III are common in hospital -65 acquired MRSA (HA-MRSA), while types IV-V are prevalent in community-acquired MRSA (CA-MRSA). 66 Livestock-associated MRSA (LA-MRSA), particularly lineage ST398, often carries SCC mec types IVa 67 or Vc (2). 68 Understanding MRSA's spread is key to combating it. S. aureus frequently colonizes human mucosal 69 surfaces, mainly the nose and the gut, where it competes with the host microbiota, especially with other 70 staphylococci (3, 4). This competition is largely influenced by nutrient availability and stressors arising 71 from the host and other bacteria, and it is still unclear how MRSA colonizes these environments (3). 72 One relevant stressor stems from bacterial mucin degradation . Here, the metabolization of sulfated 73 sugar moieties and L-cysteine residues eventually produces sulfide, which is well -known for its toxicity 74 (5). Notably, this results in high sulfide concentrations of up to 0.4 mM in the nose (6) and 0.3 - 3.4 mM 75 in the gut (7, 8) . Under oxic conditions , which are present on these mucosal surfaces (9), s ulfide 76 undergoes rapid oxidation to form inorganic polysulfides (referred to as polysulfides from here on), toxic 77 intermediates that ultimately react to form elemental sulfur (10–12). Interestingly, sulfide production and, 78 consequently, polysulfide generation are further enhanced during periods of dysbiosis and infection, 79 when the mucin layer is rapidly degraded (5). Thus, human-associated staphylococci, and in particular 80 pathogenic species like S. aureus, have to deal with these highly toxic sulfur compounds (13). 81 It is known that S. aureus possesses a sulfide-detoxifying gene cluster cst, which encodes enzymes for 82 the conversion of sulfide to less toxic sulfite (14). The cst cluster consists of genes for a sulfite transporter 83 (tauE) (15), a sulfur transferase ( cstA) (16), a sulfur dioxygenase ( cstB) (17), and a sulfide:quinone 84 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 4 oxidoreductase ( sqr) (18), which are all regulated by the transcriptional repressor CstR (19, 20) . 85 According to the current model, SQR and CstB stepwise oxidize cytoplasmic sulfide to thiosulfate (Fig. 86 S1). Then, CstA transfers the sulfane sulfur of thiosulfate to a cellular acceptor and releases sulfite, 87 which is exported out of the cell by TauE (17, 16). 88 Although the biochemical properties of the S. aureus cst gene products are well -understood, it is 89 currently unclear how other staphylococci cope with sulfide stress. Beyond that, the ecological role of 90 the cluster in staphylococc i is virtually unknown . Accordingly, the relationship between (poly -)sulfide 91 stress, detoxification, pathogenesis, and niche colonization of staphylococci, particularly MRSA, is not 92 yet understood. Strikingly, it was noted before that some MRSA strains appear to harbor a duplicate of 93 cst in proximity to the resistance-conferring mecA gene (17). However, no substantial investigation into 94 this phenomenon has been performed so far. 95

96 Distribution of the cst gene cluster in staphylococci 97 We first determined the distribution of the cst gene cluster in the genus Staphylococcus. The limited 98 annotation quality of sulfur-metabolism-associated genes and the structural similarity of respective 99 proteins despite distinct functionality (21) prompted us to use HMSS2, a recently developed tool for the 100 accurate detection of sulfur metabolism proteins (22). HMSS2 was extended to identify cst clusters in 101 proximity to mecA, with core clusters labeled cst1 and additional clusters labeled cst2. We found cst1 to 102 be widely distributed, but not highly conserved among staphylococci (Fig. 1). The species boundaries 103 roughly defined the presence of the gene cluster . However, considerable intraspecies variation was 104 observed, e.g., in S. epidermidis (16/30 strains with cst1). Notably, we found the sqr gene exclusively in 105 S. aureus and additionally on a plasmid, which is present in a few S. saprophyticus strains, raising 106 questions about its role in cst1 functionality (17). The cst2 cluster was also common in other species 107 and often SCCmec-associated, as expected from the genus-wide mobility of the genetic element (23). 108 Exceptions were the MSSA strain S. aureus ATCC 29213 and some plasmid-harboring S. saprophyticus 109 and S. pasteuri strains (Table S1 and Fig. S2) . We also found mecA-carrying strains without cst2, 110 demonstrating that not all SCCmec types harbor the cluster. 111 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 5 112 Fig 1. Distribution of cst gene clusters and mecA in the genus Staphylococcus. The cst genes and mecA 113 gene were identified using HMSS2 (22), searching the complete genome of each strain. The genes are mapped to 114 the maximum likelihood phylogeny of 229 strains from 26 species and B. subtilis 168 as outgroup. See S1 Table 115 for a comprehensive list of all strains. The cst2 genes that were found to be plasmid -located are represented as 116 unfilled circles using their respective color code. 117 Prevalence of the cst2 gene cluster on SCCmec 118 To characterize t he cst2 distribution in detail, we analyzed MRSA reference strains (24–26) and 119 epidemiologically relevant strains (27, 28), examining the spread of cst2 across various SCCmec types 120 (Table S2). We found the cst2 gene cluster in the important HA -MRSA-associated SCCmec types I-III. 121 In contrast, strains of the most prominent CA-MRSA-associated SCCmec type IV featured either no cst2 122 cluster or a truncated version (only the gene cst2B in type IVa). Furthermore, the LA-MRSA-associated 123 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 6 SCCmec type Vc strains all contained full-length cst2, while the other SCCmec subtypes Va and Vb had 124 no cst2. Additionally, SCCmec types VIII, X, XIV , and XV contained a complete cst2 cluster, whereas 125 types VI, VII, IX, and XI-XIII lacked it. In addition, we did not find any evidence for a SCCmec-located 126 sqr gene (Fig. S3). 127 We visualized the phylogenetic relationships between the genomic cst1 clusters of staphylococci and 128 the SCCmec-located cst2 clusters of MRSA (Fig. 2A). Notably, all cst2 clusters were distinct from S. 129 aureus cst1 , ruling out intraspecific duplication as the origin of the SCC mec-located clusters . 130 Furthermore, we found the SCCmec-located cst2 clusters separated into two groups: Group A cst2 from 131 SCCmec types II, III, VIII, XIV, and XV, and Group B cst2 from SCCmec types I, V, X, and the truncated 132 type IVa. Within each group, sequences are substantially conserved (≥ 99.91% identity in Group A, > 133 94.36% in Group B), but intergroup identity is approx. 72%, indicating two different origins of the clusters. 134 While some genomic cst1 clustered with Group B, and are thus potential origins of SCCmec integration, 135 we found no genomic cst1 close to Group A, rendering the origin of that cst2 currently elusive. 136 We further investigated the prevalences of the cst2 gene cluster Groups A and B within MRSA by 137 screening the NCBI complete genome database . Over one-third of all complete MRSA genomes 138 featured a full- length cst2 gene cluster , highlighting the widespread presence of cst2 among MRSA 139 strains. Out of 999 complete mecA-containing genomes (Table S3), 282 (28.23%, Table S4) contained 140 a Group A cst2 and 68 (6.81%, Table S5) contained a full -length Group B cst2 (Fig. 2A). Additionally, 141 343 (34.33%, Table S6) strains contained the truncated version of Group B cst2. 142 143 144 145 146 147 148 149 150 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 7 151 Fig 2. Genetic analysis of the cst2 cluster in MRSA. (A) Radial phylogram of representative cst1 clusters in 152 staphylococci, and all cst2 clusters identified in MRSA. The cst2 groups that were identified as conserved across 153 the diverse SCCmec types are indicated as Group A and Group B. (B) Relative abundance of Group A, full-length 154 or shortened Group B cst2 clusters (n = 282, 68, and 343, respectively) in complete MRSA genomes (n = 999) 155 found in the NCBI S. aureus complete genome database. 156 Impact of sulfide and its oxidation intermediates on S. aureus 157 The prevalence of cst2 in multiple SCCmec types suggested that MRSA benefits from the additional 158 gene cluster. The heterogeneous distribution of the cst1 gene cluster in staphylococci provided further 159 evidence that it serves as a niche adaptation with relevance to surviv al specifically in sulf ide-rich 160 environments. Thus, we investigated the ecophysiological role of the cst1 cluster in Staphylococcus 161 species to understand how an additional cst cluster may enhance MRSA fitness . Previous work had 162 shown that NaSH -containing media decreased S. aureus growth and that individual cst-encoded 163 components partially counteracted this effect (20). To study the effects of a physiologically relevant 164 sulfide concentration (1 mM) on the growth of common laboratory MSSA strain RN4220, we generated 165 a Δcst1 mutant of the strain and compared it to the wild type. Although both strains showed a prolonged 166 lag phase under 1 mM NaSH , the prolongation was substantially more pronounced in the deletion 167 mutant (Fig. 3A). Notably, a considerable increase in OD600 occurred shortly after the addition of NaSH, 168 even without bacterial cells (Fig. 3B ), suggesting a spontaneous chemical reaction. Given the oxic 169 conditions and circumneutral pH of the medium, we suspected that sulfide was being oxidized to 170 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 8 polysulfides, which would eventually react to form highly light-refracting elemental sulfur (10, 29). Due 171 to the reductive natu re of sulfide, elemental sulfur can be reduced back to polysulfides, resulting in a 172 continuous sulfide-polysulfide-sulfur interconversion (30). Additionally, sulfide can be oxidized to 173 thiosulfate and sulfite under oxic conditions (Fig. 3C). This complexity of sulfide oxidation dynamics led 174 us to identify sulfur compounds in ou r LB NaSH medium that may act as growth inhibitors. Therefore, 175 we quantified polysulfides, sulfide, thiosulfate, and sulfite by HPLC, and the insoluble elemental sulfur 176 by cyanolysis at different time points after the addition of NaSH to the medium (Fig. 3D). 177 In the first 30 minutes, a rapid decrease in sulfide concentration and a simultaneous increase in 178 polysulfide concentration was observed. After 60 minutes, sulfide was undetectable. The polysulfide 179 concentration decreased moderately as elemental sulfur was formed, and between 60 and 90 min the 180 sulfide, polysulfide, and sulfur concentrations were stable, suggesting an equilibrium in the sulfur redox 181 chemistry. Additionally, thiosulfate and sulfite were formed only in negligible amounts . Given the rapid 182 depletion of sulfide in the medium, we suspected that polysulfides, rather than sulfide, were responsible 183 for the observed growth impairment. To confirm this, we repeated the growth experiment but pre-184 incubated the LB NaSH medium for 2 h before inoculation to ensur e sulfide depletion and polysulfide 185 generation. We also included the MSSA strain Newman and its respective Δcst1 mutant (Newman 186 Δcst1) in the experiment to exclude strain-specific variations. Like in the previous experiment, produced 187 polysulfides (from 1 mM NaSH) prolonged the lag phase regardless of the strain background. Again, the 188 prolongation of the lag phase was substantially more pronounced in RN4220 Δcst1 and Newman Δcst1, 189 compared to their respective wild types (Fig. 3E and F). 190 Due to the limited prevalence of the sqr gene in the cst1 clusters of most staphylococci (and its absence 191 in cst2 clusters), a plasmid (pCQ11_cst1) was constructed that contains the cst1 cluster of S. aureus 192 Newman, purposefully excluding the sqr gene. This plasmid was introduced into RN4220 Δcst1 and 193 Newman Δcst1 to evaluate the cst-cluster's functioning without the sqr gene. The plasmid rescued the 194 wild-type behavior of both strains (RN4220 Δcst1:cst1, Newman Δcst1:cst1) in the pre sence of 195 generated polysulfides (Fig. 3E and F ), thus not only confirming the importance of the cst cluster for 196 growth under these conditions but also proving the sqr gene as non -essential for polysulfide 197 detoxification. 198 Controls with other inorganic sulfur compounds (thiosulfate, sulfite) showed no effect on growth , 199 confirming that polysulfides were the primary sulfur species affecting S. aureus growth (Fig. S4). In 200 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 9 summary, polysulfides are sufficient to cause the observed growth inhibition in S. aureus , and cst1 201 functions as a polysulfide detoxification cluster enabling faster re-initiation of bacterial growth. 202 203 Fig 3. Growth behavior of S. aureus wild type and Δcst1 mutants under NaSH/polysulfide stress. (A) Growth 204 curves of the strains RN4220 and RN4220 Δcst1 in LB NaSH (1 mM) medium. Plotted are the mean values of three 205 independent, biological replicates. Light-hued areas illustrate the standard deviations (SD) of the respective curves. 206 (B) Sulfide concentrations and OD 600nm of sterile LB NaSH (1 mM) medium over time. The mean of three 207 independent replicates is plotted . Error bars indicate the SD. (C) Schematic overview of sulfide oxidation and 208 subsequent product formation by the reaction with oxygen at near-neutral pH. (D) Formation of sulfide oxidation 209 products in sterile LB NaSH (1 mM) medium over the course of 90 min. Depicted are the mean values of three 210 independent replicates. Error bars indicate the SD. Polysulfide formation measured as ‘area under the curve’ (AUC). 211 (E) Growth curves of strains RN4220, RN4220 Δcst1, and RN4220 Δcst1:cst1 in pre-incubated LB NaSH (1 mM) 212 medium. Plotted are the mean values of three independent, biological replicates. Light -colored areas indicate the 213 SD of the respective curves. (F) Growth curves of strain Newman, Newman Δcst1, and Newman Δcst1:cst1 in pre-214 incubated LB NaSH (1 mM) medium. Data shows the mean values of three independent, biological replicates. Light-215 colored areas indicate the SD of each curve. 216 Effect of polysulfides on the growth behavior of Staphylococcus species 217 To examine wether the cst-encoded polysulfide detoxif ication is relevant to all staphylococci, we 218 measured the growth effect of generated polysulfides on multiple strains of representative 219 Staphylococcus species across the phylogeny (Fig. 4 , Fig. S 5). To determine a relative polysulfide 220 tolerance level, we used the lag phase prolongation compared to the growth b ehavior of the negative 221 control (Fig. 4A). 100% polysulfide tolerance corresponded to no difference in growth and 0% tolerance 222 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 10 indicated that we observe d no growth within 16 h ours of incubation in medium containing gene rated 223 polysulfides. Corroborating our previous results with S. aureus, we found that s trains with cst1 were 224 significantly more tolerant to polysulfides than strains without the cluster, despite some variation 225 between and within species (Fig. 4B and C ). Strains with cst1 showed moderate to high polysulfide 226 tolerances of 35.26% to 88.31%, regardless of the sqr gene. Strains lacking the cst1 gene cluster 227 generally showed little to no polysulfide tolerance. This was most noticeable in S. epidermidis and S. 228 haemolyticus, where we used wild-type strains naturally harbor ing or lacking the cluster. Overall, our 229 analysis of various staphylococci clearly showed that the presence of cst1 corresponds to higher levels 230 of polysulfide tolerance . We concluded that polysulfide-containing environments have a substantial 231 impact on staphylococcal growth and that cst1 is crucial for polysulfide detoxification to less hazardous 232 sulfite in staphylococci. 233 234 Fig 4 . Polysulfide tolerance of different staphylococci. (A) S chematic illustration depicting the process of 235 determining the relative polysulfide tolerance. The growth threshold (Gt) marks the time point where a bacterial 236 culture reached OD600nm = 0.15. Depicted are the Gt value of the growth control (Gt CTRL), the Gt of the test group 237 (GtTEST), and the difference between Gt CTRL and Gt TEST (ΔGt). (B) Relative polysulfide tolerance of different 238 staphylococci in LB containing generated polysulfides from 1 mM NaSH . Plotted are the mean values of at least 239 three independent, biological replicates. Points show the data of the individual replicates . Squares denote the 240 presence/absence of cst genes following the color scheme from Fig 1. (C) Comparison of the polysulfide tolerance 241 of the analyzed staphylococci from (B) sorted by the presence (+ cst) or absence (- cst) of cst1. Depicted are box 242 plots without outliers showing the median and the first and third quartile. Whiskers mark the minimum and maximum 243 of the data range. Asterisks indicate statistical significance (****P ≤ 0.0001) from two-tailed Student’s t-tests. Plots 244 in (B) and (C) follow the general color scheme from Fig 1 for different Staphylococcus species. 245 Polysulfide tolerance of S. aureus strains with different cst cluster configurations 246 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 11 Having determined the polysulfide detoxifying function of the cst1 cluster products, we investigated how 247 SCCmec-carrying sta phylococci benefit from an additional cst2 cluster. First, we complemented the 248 MSSA mutant strains RN4220 Δcst1 and Newman Δcst1, with a plasmid (pCQ11_cst2) carrying the cst2 249 from the MRSA strain COL. This rescued the wild-type behavior in both strains, verifying that cst2 is 250 fully functioning as a polysulfide detoxification cluster (Fig. S6). Notably, the Δcst1:cst2 strains behaved 251 similarly to the Δcst1:cst1 strains, illustrating that the gene clusters are functionally synonymous. Thus, 252 we suspected that the additional cst2 might further improve polysulfide tolerance in MRSA compared to 253 strains carrying only one cst cluster. Therefore, we tested the polysulfide tolerance of various MRSA 254 and MSSA strains with different cst cluster configurations (Fig. 5, Fig. S7A). In general, w e observed 255 high polysulfide tolerance in S. aureus at low polysulfide concentrations (from 1 mM NaSH), independent 256 of the strain background (MRSA or MSSA). Two MRSA strains naturally lacked cst clusters and showed 257 no polysulfide tolerance , corroborating our previous results . With polysulfides generated from 3 mM 258 NaSH, however, the polysulfide tolerance of the MSSA strains decreased significantly (Fig. 5A and B). 259 Remarkably, MRSA strains still had high tolerance levels under these conditions, resulting in a 260 significant difference between strains carrying one or two cst clusters (Fig. 5A and B). 261 We hypothesized that the cst2 cluster may improve MRSA fitness at higher polysulfide concentratio ns 262 or in the presence of polysulfides with longer chain lengths . To this end , w e repeated the previous 263 experiment using polysulfides (1 mM) with defined chain lengths (Na2S2, Na2S3, and Na2S4). Now, the 264 difference between MSSA and MRSA became even clearer (Fig. 5A, Fig. S7B). Again, strains lacking 265 the cst cluster could not grow. The MSSA strains showed a heterogeneous behavior in the Na2S2 setup, 266 with half the strains exhibiting medium tolerance, while the others failed to thrive. With increasing chain 267 lengths MSSA tolerance steadily declined. In the Na 2S4 setup, MSSA polysulfide tolerance was 268 drastically limited, with only one of eight strains showing more than 10% tolerance. In contrast, MRSA 269 strains had significantly higher tolerance levels to all thr ee polysulfide chain lengths tested. We still 270 observed that increasing the chain length affected MRSA tolerance to a varying degree. However, the 271 mean tolerance of MRSA strains consistently exceeded that of the most tolerant MSSA strain. Even in 272 the presence of 1 mM Na2S4, all MRSA strains grew. None showed less than 10% tolerance and two of 273 six strains still exhibited over 50% tolerance. 274 To validate these results, we replicated the experiment with genetically modified strain s, introducing a 275 plasmid-located cst2 (pCQ11_cst2) from COL into MSSA wild-type strains (RN4220 and Newman), and 276 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 12 deleting cst1 from MRSA COL (Fig. 5C). While the polysulfides generated from 1 mM or 3 mM NaSH 277 induced only small, non-significant differences between the strains harboring one or two cst clusters 278 (Fig. 5B), significant differences were observed when 1 mM Na2S3 and 1 mM Na2S4 were applied (Fig. 279 5D). Strains having both clusters ( cst1 and cst2) were significantly more tolerant than strains with only 280 one cluster at polysulfide chain lengths of 3 or 4, regardless of the strain background and cst-cluster-281 configuration. Consequently, MSSA mutants with added cst2 behaved similarly to MRSA wild-type 282 strains (see also Fig. 5A for comparison) , while the MRSA mutant COL carrying only cst2 was 283 phenotypically comparable to the MSSA wild -type strains. These results confirmed that the additional 284 SCCmec-located cst2 is largely responsible for the phenotypic polysulfide tolerance in MRSA. Moreover, 285 the use of polysulfide standards demonstrated that the inhibitory effect is not entirely dependent on the 286 polysulfide concentration but rather on the amount of sulfur bound in the reactive polysulfide form. 287 288 Fig 5. Impact of polysulfides on different S. aureus strains. (A) Heat maps displaying the relative polysulfide 289 tolerance of multiple S. aureus wild-type strains at different polysulfide concentrations (“low”, generated from 1 mM 290 NaSH; “high”, generated from 3 mM NaSH ) and chain lengths ( 1 mM of Na2S2, Na 2S3, or Na2S4) after 16 h of 291 incubation. Depicted are the mean values of at least three independent, biological replicates. Squares and circles 292 indicate the presence/absence of cst1 or cst2 genes, following the color scheme from Fig 1. MRSA strains are 293 marked with asterisks in front of their strain designation. (B) Relative polysulfide tolerance of the tested strains from 294 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 13 (A) grouped by harboring cst1 + cst2 (n=6), only cst1 (n=9), or none (n=2). The box plots show the median and the 295 first and third quartiles. Whiskers mark the minimum and maximum of the data range. Asterisks indicate statistical 296 significance (ns = not significant; *P = 0.05 to 0.01; ** P = 0.01 to 0.001; *** P = 0.001 to 0.0001; **** P ≤ 0.0001) 297 from two-tailed Student’s t-tests. (C) Heat maps displaying the relative polysulfide tolerance of S. aureus cst-mutant 298 strains and their respective wild types at different polysulfide concentrations ( “low”, generated from 1 mM NaSH; 299 “high”, generated from 3 mM NaSH) and chain lengths (1 mM of Na2S2, Na2S3, or Na2S4) after 16 h of incubation. 300 Plotted are the mean values of at least three independent, biological replicates. Squares and circles indicate the 301 presence/absence of cst1 or cst2 genes, following the color scheme from Fig 1. MRSA strains are marked with 302 asterisks in front of their strain designation. (D) Relative polysulfide tolerance of the tested strains from (C) grouped 303 by their cst configuration (cst1 + cst2 n=3, cst1 n=3, and ∆cst1 n=2) and depicted as box plots. Shown are the 304 median and first and third quartiles. Whiskers mark the minimum and maximum of the data range. Statistical 305 significance from unpaired two-tailed Student’s t-tests is denoted as asterisks (ns = not significant; *P = 0.05 to 306 0.01; **P = 0.01 to 0.001; ***P = 0.001 to 0.0001). 307 Intraspecies competition between MSSA and MRSA under polysulfide stress 308 Based on the reported results, we hypothesized that the increased polysulfide tolerance of MRSA may 309 enhance its competitive fitness against MSSA in polysulfide-containing environments. Therefore, we co-310 incubated MSSA strain RN4220 with either of two MRSA strains (COL and N315, representing SCCmec 311 type I with Group B cst2, and SCCmec type II with Group A cst2, respectively) in a 99:1 OD600 ratio. We 312 monitored their relative abundance over three days, with daily passaging into fresh medium containing 313 1 mM Na 2S3. Without polysulfides, the MSSA strain remained at a n overwhelming majority of ≥ 99% 314 relative abundance throughout the experiment (Fig. 6), indicating that the MRSA failed to outcompete 315 the other strain and invade the occupied environment. However, under polysulfide stress conditions, the 316 MRSA strains were able to compete with the MSSA strain RN4220, establishing a substantial 317 subpopulation of over 15% within the first 24 h of co-incubation (Fig. 6). After 48 hours, the MRSA strains 318 dominated the co-cultures. After 72 hours, both MRSA strains represented over 98% of the co-cultures, 319 completely reversing the initial ratio within three days. As expected, the MSSA strain Newman could not 320 benefit from the polysulfide stress, and the strain was thus unable to outcompete the occupying RN4220 321 during the control experiment. This demonstrated that the SCCmec-located cst2 enables MRSA to 322 drastically outcompete MSSA in polysulfide-rich environments. 323 324 325 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 14 326 Fig 6. Competitive growth behavior of MRSA and MSSA (Newman) against the MSSA strain RN4220. Shown 327 are relative strain abundances during serial passaging over the course o f 72 h in medium with or without 1 mM 328 Na2S3. Depicted are the mean values of three independent, biological replicates. Error bars indicate SD. Statistical 329 significance from unpaired two-tailed Student’s t-tests is denoted as asterisks ( ns = not significant; *P = 0.05 to 330 0.01; **P = 0.01 to 0.001; ***P = 0.001 to 0.0001; ****P ≤ 0.0001). 331

332 In this work, we have elucidated the ecophysio logical role of the cst gene clusters in mediating 333 polysulfide detoxification in staphylococci. Notably, we have shown that MRSA strains frequently harbor 334 an additional, SCCmec-located cst cluster, namely cst2, which substantially increases polysulfide 335 tolerance and provides deciding competitive fitness in polysulfide-rich environments. 336 The widespread, heterogeneous distribution of the cst1 gene cluster in staphylococci emphasizes its 337 importance as an adaptation to specific, sulfide-containing habitats. Although Staphylococcus species 338 are common colonizers of the human body, not all host-associated staphylococci will encounter (poly-339 )sulfides, as the conditions necessary for generating these compounds differ drastically among different 340 body sites. While strains inhabiting mucosal surfaces presumably face these stressors regularly, other 341 strains colonizing, e.g., the skin are much less exposed to (poly -)sulfides. Thus, possessing cst gene 342 clusters may serve as an adaptation to mucosal habitats. Moreover, since various forms of inflammation 343 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 15 are linked to increased sulfide levels (31), the cst clusters may be crucial for pathogenic or infection -344 associated strains. 345 In contrast to the overall abundance of the cst1 cluster, the sqr gene was predominantly restricted to S. 346 aureus and was never present in cst2 clusters. This suggests a particular ecological function of the SQR 347 beyond the general role of cst1. In support of this, we observed that cst clusters provide fully functional 348 polysulfide detoxification without sqr, demonstrating that TauE, CstA, and CstB are sufficient to confer 349 polysulfide tolerance to staphylococci. Thus, we propose to include polysulfide detoxification without 350 SQR as an extension to the current model of the gene cluster’s function. Without the need for the initial 351 sulfide oxidation by SQR, polysulfide s may directly interact with CstA, which transfers individual thiol 352 groups onto a suitable cellular sulfur acceptor, i.e. a low-molecular-weight thiol (LMW-SH). The resulting 353 LMW-SSH may be handled by CstB, which oxidizes the persulfide sulfur with the consumption of O 2 354 and releases sulfite, which is finally transported out of the cell by TauE (Fig. S8). Longer polysulfide 355 chains may increase stress due to the increased amount of reactive thiol groups that need to be 356 transferred to CstA , which is supported by our results with the polysulfide standards. Regarding the 357 ecological function of SQR in S. aureus, it remains elusive how its well-characterized role in the sulfide-358 to-thiosulfate conversion with concomitant quinone reduction (18) translates into a trait that is beneficial 359 for this species. 360 While great advances have been made to elucidate various effects of H 2S on bacterial physiology, little 361 is known about the impact of other reactive sulfur compounds (32). Our results demonstrate that 362 polysulfides are highly toxic to staphylococci, matching or exceeding H2S toxicity. Detoxification by cst1-363 encoded enzymes reduces exposure time to harmful polysulfide concentrations , contributing to the 364 observed polysulfide tolerance. The SCCmec-located cst2 cluster provides additional detoxification 365 capacity, enhancing the tolerance to high polysulfide stress. Our bioinformatics analysis demonstrated 366 that cst2 is widespread amongst MRSA strains , spanning key SCCmec types and multiple 367 epidemiologically relevant strain s (27, 28). To the best of our knowledge, this constitutes cst2 as the 368 most abundant SCCmec-located gene cluster (aside from the SCCmec-defining mec and ccr 369 complexes). This strongly suggests that MRSA benefit s from cst2-conferred tolerance to proliferate in 370 (poly-)sulfide-rich niches, like the nose and, in particular, the gut (33, 34, 4) . Moreover, this may be 371 especially important during events of dysbiosis and inflammation, where levels of reactive sulfur species 372 are elevated (5). In turn, this may be relevant in clinical settings, where prolonged hospitalizations and 373 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 16 medical predispositions often cause steady levels of dysbiosis and inflammation (e.g., long-term care 374 facilities, dialysis, post-surgery, and ICU admission) (35, 36). In support, we found that the cst2 cluster 375 is prominent in SCCmec types strongly associated with HA-MRSA (27, 28). 376 Since our co-incubation experiments demonstrated that the cst2 cluster effectively provides MRSA with 377 a substantial advantage in intraspecies competition against already established MSSA populations 378 during polysulfide stress, the additional gene cluster might act as a colonization factor enabling MRSA 379 to invade the human microbiome. This might be especially relevant because the microbiome of mucosal 380 surfaces itself is a constant source of considerable (poly-)sulfide concentrations (6, 37, 7, 8, 38) . 381 Polysulfide tolerance may therefore select for MRSA over less tolerant staphylococci, particularly during 382 states of disease. 383 Overall, our work shows that polysulfide stress is a biologically relevant factor for staphylococci that may 384 contribute to the spread of cst2-containing SCCmec and MRSA. Thus, we suggest that the SCCmec-385 located cst2 cluster is a critical factor facilitating MRSA proliferation within the human microbiome. 386

387 Bioinformatics analyses 388 The cst gene cluster was identified using a modified version of the HMSS2 tool (22) by extending the 389 library to include the proteins CstR, CstA and MecA. The Distribution of the cst gene cluster and mecA 390 were visualized in a maximum likelihood phylogeny . For detailed information, see supplementa ry 391

and methods. 392 The presence of cst2 in SCCmec cassettes was determined using BLAST for a list of representative 393 MRSA strains (Table S2), compiled from sources listing reference strains for SCC mec typing (24–26) 394 and epidemiologically relevant strains (27, 41, 28) . To generate the phylogram of staphylococcal cst 395 gene clusters, all thereby found SCCmec-located cst2 clusters and representative genomic 396 staphylococcal cst1 gene clusters detected with the HMSS2 tool were aligned using EMBL-EBI Clustal 397 Omega (42). The phylogenetic tree was calculated using IQTree 1.6.12 (43) and visualized with FigTree 398 v1.4.4. Representative SCCmec cassettes carrying cst2 were aligned using EMBL-EBI Clustal Omega 399 (42) and visualized using clinker (44). 400 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 17 To determine cst2 distribution in publicly available complete MRSA genomes, cst2 from N315 or COL 401 were used as query sequences for cst2 Group A and B, respectively, in a BLAST genome search of the 402 NCBI S. aureus (taxid:1280) complete genome database . Results were filtered for > 90% identity to 403 exclude S. aureus cst1 and the respective other cst2 group. Shortened G roup B cst2 (cstB2) was 404 identified by query cover. The total number of MRSA genomes was determined by a similar BLAST 405 genome search using S. aureus N315 mecA as query se quence and all strains featuring a full -length 406 cst2 were filtered for the presence of mecA by list comparison with the results of this search. 407 Strains and plasmids 408 All primers, plasmids, and strains used in this study are listed in Table S7-9. The generation of S. aureus 409 ∆cst1 mutants and the introduction of plasmid-located cst1 and cst2 into S. aureus, as well as the 410 detection of the cst gene cluster in Staphylococcus spp. are described in supplementary material and 411 methods. 412 Standard growth conditions 413 All strains were cultivated in Lysogeny Broth (LB, 10 g/l NaCl) and incubated at 37 °C and 120 rpm 414 agitation, if not stated otherwise. 415 Generation of polysulfides in LB medium 416 To generate different concentrations of polysulfides (low, high), LB medium was supplemented with 417 different concentrations of NaSH (1 mM and 3 mM, respectively) and incubated for 2 h at 37 °C with 418 120 rpm agitation to ensure sufficient polysulfide generation. 419 420 Growth experiments with sulfide and polysulfides 421 To study the impact of sulfur species on the growth of different staphylococci, 99 µl of generated 422 polysulfides (see above), 1 mM NaSH, 1 mM Na 2Sn standards (Na2S2, Na2S3 or Na2S4), 1 mM S2O32−, 423 1 mM SO 32−, or 1 mM SO 42− dissolved in LB medium were transferred into a 96-well plate, inoculated 424 with 1%(v/v) of an overnight grown culture (OD600nm adjusted to 0.3 prior to inoculation), and sealed with 425 gas-permeable MICRONAUT sealing foil (sifin diagnostic). The optical density was measured every 5 426 min for 16 h at 37 °C using a Tecan infinite 200 PRO (Tecan). 427 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 18 The lag phase prolongation was used to determine a relative polysulfide tolerance level compared to 428 the growth behavior of the negative control. 100% polysulfide tolerance corresponded to no difference 429 in growth and 0% tolerance denoted that we did not observe growth within 16 h of incubation in medium 430 containing generated polysulfides . To de termine the time point where cu ltures re -initiated growth, a 431 growth threshold (Gt) value at which cultures reached OD600nm = 0.15 (marking the beginning of the log 432 phase) was defined (see also Fig. 4A ). The difference (ΔGt) between the Gt value under polysulfide 433 stress (GtTEST) and the Gt value of the negative control (GtCTRL) was then used to calculate the polysulfide 434 tolerance using equation (1). 435 𝑝𝑜𝑙𝑦𝑠𝑢𝑙𝑓𝑖𝑑𝑒 𝑡𝑜𝑙𝑒𝑟𝑎𝑛𝑐𝑒 [%] = 100% − 100% × ∆Gt [min] 𝑡𝑜𝑡𝑎𝑙 𝑔𝑟𝑜𝑤𝑡ℎ 𝑡𝑖𝑚𝑒 [min] (1) 436 Competition Assay 437 Overnight cultures of MSSA (RN4220, Newman) and MRSA (COL, N315) strains were adjusted to OD600 438 = 1.5 and mixed at an OD600 ratio of 99:1 (MSSA/MRSA, MSSA/MSSA) to a final volume of 1.5 ml. LB 439 medium with or without 1 mM Na2S3 was inoculated with 1% (v/v) of the mixture and incubated for 24 h 440 at 37 °C with 120 rpm agitation. After 24 h and 48 h, fresh medium with or without 1 mM Na 2S3 was 441 inoculated with 1% (v/v) of the competition mixture. For each time point (0 h, 24 h, 48 h, 72 h), the 442 competition mixture was serially diluted in 0.9% NaCl solution and 10µl of it was spotted on LB plates 443 for total cell numbers and LB plates with antibiotic for strain-specific selection (10 µg/ml erythromycin, 444 2.5 µg/ml oxacillin, or 10 µg/ml chloramphenicol, respectively, see Table S11). After 24 h of growth, the 445 colony-forming units (CFUs) per ml were determined for each strain, and the MSSA/MRSA and 446 MSSA/MSSA ratios were determined. 447 Determination of sulfur species 448 To determine the concentrations of HS- and its oxidation products in medium over time, LB medium was 449 substituted with 1 mM NaSH and incubated for 2 h at 37 °C with 120 rpm agitation. At different time 450 points, samples were taken to quantify the sulfide species. Thioles were quantified after derivatization 451 with the fluorescent dye monobromobimane as previously described (45). S 8 was colorimetrically 452 determined after treatment with cyanide as described elsewhere (46). See supplementary material and 453

for a detailed description. 454 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 19 Statistical analysis 455 Statistical analysis of data was performed using GraphPad Prism 10.2.0. If not stated otherwise, 456 statistical significance was determined in unpaired two- tailed Student’s t-tests with a 95% confidence 457 interval against the respective standard or negative control. All experiments were repeated at least three 458 times with biologically independent replicates. All graphs show the mean of at least three biological 459 replicates with error bars denoting the SD. Statistical significance was denoted as not significant (ns); 460 *P=0.05 to 0.01; **P=0.01 to 0.001; ***P=0.001 to 0.0001; ****P≤0.0001. 461 Data availability: 462 All raw data for graphs are supplied in source data files. Source data are provided with this paper. 463 Code availability: 464 HMSS2 program files are available at https://github.com/TSTanabe/HMSS2. 465

466 We thank Jana Rohe and Kainat Qureshi for technical assistance, and Prof. Dr. Gabriele Bierbaum and 467 her research group for supplying us with staphylococcal isolates from their remarkable collection. We 468 would also like to thank Dr. Stefania De Benedetti for additional proofreading. This work was mostly 469 funded by the Jürgen Manchot Foundation and the German Center for Infection Research (DZIF) . 470 Tomohisa Sebastian Tanabe acknowledges funding from the Austrian Science Fund (FWF) 471 [doi.org/10.55776/COE7]. 472 Ethics declarations: 473 The authors declare no competing interests. 474

475 1. Murray CJL, Ikuta KS, Sharara F, Swetschinski L, Robles Aguilar G, Gray A, Han C, 476 Bisignano C, Rao P, Wool E, Johnson SC, Browne AJ, Chipeta MG, Fell F, Hackett S, 477 Haines-Woodhouse G, Kashef Hamadani BH, Kumaran EAP, McManigal B, Agarwal R, 478 Akech S, Albertson S, Amuasi J, Andrews J, Aravkin A, Ashley E, Bailey F, Baker S, 479 Basnyat B, Bekker A, Bender R, Bethou A, Bielicki J, Boonkasidecha S, Bukosia J, 480 Carvalheiro C, Castañeda-Orjuela C, Chansamouth V, Chaurasia S, Chiurchiù S, 481 Chowdhury F, Cook AJ, Cooper B, Cressey TR, Criollo-Mora E, Cunningham M, Darboe 482 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 20 S, Day NPJ, Luca M de, Dokova K, Dramowski A, Dunachie SJ, Eckmanns T, Eibach D, 483 Emami A, Feasey N, Fisher-Pearson N, Forrest K, Garrett D, Gastmeier P, Giref AZ, 484 Greer RC, Gupta V, Haller S, Haselbeck A, Hay SI, Holm M, Hopkins S, Iregbu KC, 485 Jacobs J, Jarovsky D, Javanmardi F, Khorana M, Kissoon N, Kobeissi E, Kostyanev T, 486 Krapp F, Krumkamp R, Kumar A, Kyu HH, Lim C, Limmathurotsakul D, Loftus MJ, Lunn 487 M, Ma J, Mturi N, Munera-Huertas T, Musicha P, Mussi-Pinhata MM, Nakamura T, 488 Nanavati R, Nangia S, Newton P, Ngoun C, Novotney A, Nwakanma D, Obiero CW, 489 Olivas-Martinez A, Olliaro P, Ooko E, Ortiz-Brizuela E, Peleg AY, Perrone C, Plakkal N, 490 Ponce-de-Leon A, Raad M, Ramdin T, Riddell A, Roberts T, Robotham JV, Roca A, 491 Rudd KE, Russell N, Schnall J, Scott JAG, Shivamallappa M, Sifuentes-Osornio J, 492 Steenkeste N, Stewardson AJ, Stoeva T, Tasak N, Thaiprakong A, Thwaites G, Turner 493 C, Turner P, van Doorn HR, Velaphi S, Vongpradith A, Vu H, Walsh T, Waner S, 494 Wangrangsimakul T, Wozniak T, Zheng P, Sartorius B, Lopez AD, Stergachis A, Moore 495 C, Dolecek C, Naghavi M. 2022. Global burden of bacterial antimicrobial resistance in 496 2019: a systematic analysis. The Lancet 399:629–655. 497 2. Fernandez JE, Egli A, Overesch G, Perreten V. 2024. Time-calibrated phylogenetic and 498 chromosomal mobilome analyses of Staphylococcus aureus CC398 reveal geographical 499 and host-related evolution. Nat Commun 15:5526. 500 3. Piewngam P, Otto M. 2024. Staphylococcus aureus colonisation and strategies for 501 decolonisation. The Lancet Microbe 5:e606-e618. 502 4. Piewngam P, Khongthong S, Roekngam N, Theapparat Y, Sunpaweravong S, 503 Faroongsarng D, Otto M. 2023. Probiotic for pathogen-specific Staphylococcus aureus 504 decolonisation in Thailand: a phase 2, double-blind, randomised, placebo-controlled 505 trial. The Lancet Microbe 4:e75-e83. 506 5. Stummer N, Feichtinger RG, Weghuber D, Kofler B, Schneider AM. 2023. Role of 507 Hydrogen Sulfide in Inflammatory Bowel Disease. Antioxidants (Basel) 12. 508 6. Ikeda M, Ishima Y, Chuang VTG, Sakai M, Osafune H, Ando H, Shimizu T, Okuhira K, 509 Watanabe H, Maruyama T, Otagiri M, Akaike T, Ishida T. 2019. Distribution of 510 Polysulfide in Human Biological Fluids and Their Association with Amylase and Sperm 511 Activities. Molecules 24. 512 7. Braccia DJ, Jiang X, Pop M, Hall AB. 2021. The Capacity to Produce Hydrogen Sulfide 513 (H2S) via Cysteine Degradation Is Ubiquitous in the Human Gut Microbiome. Front 514 Microbiol 12:705583. 515 8. Dordević D, Jančíková S, Vítězová M, Kushkevych I. 2021. Hydrogen sulfide toxicity in 516 the gut environment: Meta-analysis of sulfate-reducing and lactic acid bacteria in 517 inflammatory processes. J Adv Res 27:55–69. 518 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 21 9. Espey MG. 2013. Role of oxygen gradients in shaping redox relationships between the 519 human intestine and its microbiota. Free Radic Biol Med 55:130–140. 520 10. Chen KY, Morris JC. 1972. Kinetics of oxidation of aqueous sulfide by oxygen. Environ. 521 Sci. Technol. 6:529–537. 522 11. Kimura Y, Mikami Y, Osumi K, Tsugane M, Oka J, Kimura H. 2013. Polysulfides are 523 possible H2S-derived signaling molecules in rat brain. FASEB J 27:2451–2457. 524 12. Olson KR. 2012. A practical look at the chemistry and biology of hydrogen sulfide. 525 Antioxid Redox Signal 17:32–44. 526 13. Jiang J, Chan A, Ali S, Saha A, Haushalter KJ, Lam W-LM, Glasheen M, Parker J, 527 Brenner M, Mahon SB, Patel HH, Ambasudhan R, Lipton SA, Pilz RB, Boss GR. 2016. 528 Hydrogen Sulfide--Mechanisms of Toxicity and Development of an Antidote. Sci Rep 529 6:20831. 530 14. Han X, Zhu F, Chen L, Wu H, Wang T, Chen K. 2020. Mechanism analysis of toxicity of 531 sodium sulfite to human hepatocytes L02. Mol Cell Biochem 473:25–37. 532 15. Weinitschke S, Denger K, Cook AM, Smits THM. 2007. The DUF81 protein TauE in 533 Cupriavidus necator H16, a sulfite exporter in the metabolism of C2 sulfonates. 534 Microbiology (Reading) 153:3055–3060. 535 16. Higgins KA, Peng H, Luebke JL, Chang F-MJ, Giedroc DP. 2015. Conformational 536 analysis and chemical reactivity of the multidomain sulfurtransferase, Staphylococcus 537 aureus CstA. Biochemistry 54:2385–2398. 538 17. Shen J, Keithly ME, Armstrong RN, Higgins KA, Edmonds KA, Giedroc DP. 2015. 539 Staphylococcus aureus CstB is a novel multidomain persulfide dioxygenase-540 sulfurtransferase involved in hydrogen sulfide detoxification. Biochemistry 54:4542–541 4554. 542 18. Shen J, Peng H, Zhang Y, Trinidad JC, Giedroc DP. 2016. Staphylococcus aureus sqr 543 Encodes a Type II Sulfide:Quinone Oxidoreductase and Impacts Reactive Sulfur 544 Speciation in Cells. Biochemistry 55:6524–6534. 545 19. Grossoehme N, Kehl-Fie TE, Ma Z, Adams KW, Cowart DM, Scott RA, Skaar EP, 546 Giedroc DP. 2011. Control of Copper Resistance and Inorganic Sulfur Metabolism by 547 Paralogous Regulators in Staphylococcus aureus*. J Biol Chem 286:13522–13531. 548 20. Luebke JL, Shen J, Bruce KE, Kehl-Fie TE, Peng H, Skaar EP, Giedroc DP. 2014. The 549 CsoR-like sulfurtransferase repressor (CstR) is a persulfide sensor in Staphylococcus 550 aureus. Mol Microbiol 94:1343–1360. 551 21. Tanabe TS, Dahl C. 2022. HMS-S-S: A tool for the identification of Sulphur metabolism-552 related genes and analysis of operon structures in genome and metagenome 553 assemblies. Mol Ecol Resour 22:2758–2774. 554 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 22 22. Tanabe TS, Dahl C. 2023. HMSS2: An advanced tool for the analysis of sulphur 555 metabolism, including organosulphur compound transformation, in genome and 556 metagenome assemblies. Mol Ecol Resour 23:1930–1945. 557 23. Garza-González E, Morfín-Otero R, Llaca-Díaz JM, Rodriguez-Noriega E. 2010. 558 Staphylococcal cassette chromosome mec (SCCmec) in methicillin-resistant coagulase-559 negative staphylococci. A review and the experience in a tertiary-care setting. Epidemiol 560 Infect 138:645–654. 561 24. Kaya H, Hasman H, Larsen J, Stegger M, Johannesen TB, Allesøe RL, Lemvigh CK, 562 Aarestrup FM, Lund O, Larsen AR. 2018. SCCmecFinder, a Web-Based Tool for Typing 563 of Staphylococcal Cassette Chromosome mec in Staphylococcus aureus Using Whole-564 Genome Sequence Data. mSphere 3. 565 25. Uehara Y. 2022. Current Status of Staphylococcal Cassette Chromosome mec 566 (SCCmec). Antibiotics (Basel) 11. 567 26. 2009. Classification of staphylococcal cassette chromosome mec (SCCmec): guidelines 568 for reporting novel SCCmec elements. Antimicrob Agents Chemother 53:4961–4967. 569 27. Lakhundi S, Zhang K. 2018. Methicillin-Resistant Staphylococcus aureus: Molecular 570 Characterization, Evolution, and Epidemiology. Clin Microbiol Rev 31. 571 28. Jiang J-H, Cameron DR, Nethercott C, Aires- de-Sousa M, Peleg AY. 2023. Virulence 572 attributes of successful methicillin-resistant Staphylococcus aureus lineages. Clin 573 Microbiol Rev 36:e0014822. 574 29. Steudel R. 1996. Mechanism for the Formation of Elemental Sulfur from Aqueous 575 Sulfide in Chemical and Microbiological Desulfurization Processes. Ind. Eng. Chem. 576 Res. 35:1417–1423. 577 30. Kamyshny A, Goifman A, Gun J, Rizkov D, Lev O. 2004. Equilibrium distribution of 578 polysulfide ions in aqueous solutions at 25 degrees C: a new approach for the study of 579 polysulfides' equilibria. Environ. Sci. Technol. 38:6633–6644. 580 31. Dilek N, Papapetropoulos A, Toliver-Kinsky T, Szabo C. 2020. Hydrogen sulfide: An 581 endogenous regulator of the immune system. Pharmacol Res 161:105119. 582 32. Greiner R, Pálinkás Z, Bäsell K, Becher D, Antelmann H, Nagy P, Dick TP. 2013. 583 Polysulfides link H2S to protein thiol oxidation. Antioxid Redox Signal 19:1749–1765. 584 33. Acton DS, Plat-Sinnige MJT, van Wamel W, Groot N de, van Belkum A. 2009. Intestinal 585 carriage of Staphylococcus aureus: how does its frequency compare with that of nasal 586 carriage and what is its clinical impact? Eur J Clin Microbiol Infect Dis 28:115–127. 587 34. Turner NA, Sharma-Kuinkel BK, Maskarinec SA, Eichenberger EM, Shah PP, Carugati 588 M, Holland TL, Fowler VG. 2019. Methicillin-resistant Staphylococcus aureus: an 589 overview of basic and clinical research. Nat Rev Microbiol 17:203–218. 590 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 23 35. Cho NA, Strayer K, Dobson B, McDonald B. 2024. Pathogenesis and therapeutic 591 opportunities of gut microbiome dysbiosis in critical illness. Gut Microbes 16:2351478. 592 36. Kullberg RFJ, Wikki I, Haak BW, Kauko A, Galenkamp H, Peters-Sengers H, Butler JM, 593 Havulinna AS, Palmu J, McDonald D, Benchraka C, Abdel-Aziz MI, Prins M, van der 594 Maitland Zee AH, van den Born B-J, Jousilahti P, Vos WM de, Salomaa V, Knight R, 595 Lahti L, Nieuwdorp M, Niiranen T, Wiersinga WJ. 2024. Association between butyrate-596 producing gut bacteria and the risk of infectious disease hospitalisation: results from two 597 observational, population-based microbiome studies. The Lancet Microbe 5:100864. 598 37. Baskar R, Bian J. 2011. Hydrogen sulfide gas has cell growth regulatory role. Eur J 599 Pharmacol 656:5–9. 600 38. Furne J, Saeed A, Levitt MD. 2008. Whole tissue hydrogen sulfide concentrations are 601 orders of magnitude lower than presently accepted values. Am J Physiol Regul Integr 602 Comp Physiol 295:R1479-85. 603 39. Hyatt D, Chen G-L, Locascio PF, Land ML, Larimer FW, Hauser LJ. 2010. Prodigal: 604 prokaryotic gene recognition and translation initiation site identification. BMC 605 Bioinformatics 11:119. 606 40. Letunic I, Bork P. 2021. Interactive Tree Of Life (iTOL) v5: an online tool for phylogenetic 607 tree display and annotation. Nucleic Acids Res 49:W293-W296. 608 41. Liu J, Chen D, Peters BM, Li L, Li B, Xu Z, Shirliff ME. 2016. Staphylococcal 609 chromosomal cassettes mec (SCCmec): A mobile genetic element in methicillin-610 resistant Staphylococcus aureus. Microb Pathog 101:56–67. 611 42. Madeira F, Pearce M, Tivey ARN, Basutkar P, Lee J, Edbali O, Madhusoodanan N, 612 Kolesnikov A, Lopez R. 2022. Search and sequence analysis tools services from EMBL-613 EBI in 2022. Nucleic Acids Res 50:W276-W279. 614 43. Nguyen L-T, Schmidt HA, Haeseler A von, Minh BQ. 2015. IQ-TREE: a fast and 615 effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol Biol 616 Evol 32:268–274. 617 44. Gilchrist CLM, Chooi Y-H. 2021. clinker & clustermap.js: automatic generation of gene 618 cluster comparison figures. Bioinformatics 37:2473–2475. 619 45. Rethmeier J, Rabenstein A, Langer M, Fischer U. 1997. Detection of traces of oxidized 620 and reduced sulfur compounds in small samples by combination of different high-621 performance liquid chromatography methods. Journal of Chromatography A 760:295–622 302. 623 46. Dahl C. 1996. Insertional gene inactivation in a phototrophic sulphur bacterium: APS-624 reductase-deficient mutants of Chromatium vinosum. Microbiology (Reading) 142 (Pt 625 12):3363–3372. 626 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 24 627 Figure 2: Distribution of cst gene clusters and mecA in the genus Staphylococcus, mapped to the maximum 628 likelihood phylogeny of 229 strains from 26 species and B. subtilis 168 as outgroup. The cst genes and mecA 629 gene were identified using HMSS2 (22), searching the complete genome of each strain. See Supplementary Tab. 630 S1 for a comprehensive list of all strains. The cst2 genes that were found to be plasmid-located are represented as 631 unfilled circles using their respective color code. 632 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 25 633 Figure 2: Genetic analysis of the cst2 cluster in MRSA. A) Radial phylogram of representative cst1 clusters in 634 staphylococci, and all cst2 clusters identified in MRSA. The cst2 groups that were identified as conserved across 635 the diverse SCCmec types are indicated as Group A and Group B. B) Relative abundance of Group A, full-length 636 or shortened Group B cst2 clusters (n = 282, 68, and 343, re spectively) in complete MRSA genomes (n = 999) 637 found in the NCBI S. aureus complete genome database. 638 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 26 639 Figure 3: Growth behavior of S. aureus wild type and Δcst1 mutants in LB medium supplemented with 1 640 mM NaSH/generated polysulfides. A) Growth curves of the strains RN4220 and RN4220 Δcst1 in LB NaSH (1 641 mM) medium. Plotted are the mean values of three independent, biological replicates. Light -hued areas illustrate 642 the standard deviations (SD) of the respective curves. B) Sulfide concentrations and OD600nm of sterile LB NaSH (1 643 mM) medium over time. The mean of three independent replicates is plotted. E rror bars indicate the SD. C ) 644 Schematic overview of sulfide oxidation and subsequent product formation by the reaction w ith oxygen at near-645 neutral pH. D ) Formation of sulfide oxidation products in sterile LB NaSH (1 mM) medium over the course of 90 646 min. Depicted are the mean values of three independent replicates. Error bars indicate the SD. Polysulfide formation 647 measured as ‘area under the curve’ (AUC). E) Growth curves of strains RN4220, RN4220 Δcst1, and RN4220 648 Δcst1:cst1 in pre-incubated LB NaSH (1 mM) medium. Plotted are the mean values of three independent, biological 649 replicates. Light -colored areas indicate th e SD of the respective curves. F ) Growth curves of strain Newman, 650 Newman Δcst1, and Newman Δcst1:cst1 in pre-incubated LB NaSH (1 mM) medium. Data shows the mean values 651 of three independent, biological replicates. Light-colored areas indicate the SD of each curve. 652 653 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 27 654 Figure 4: Polysulfide tolerance of different staphylococci. A) Schematic illustration depicting the process of 655 determining the relative polysulfide tolerance. The growth threshold (Gt) marks the time point where a bacterial 656 culture reached OD600nm = 0.15. Depicted are the Gt value of the growth control (Gt CTRL), the Gt of the test group 657 (GtTEST), and the difference between Gt CTRL and Gt TEST (ΔGt). B ) Relative polysulfide tolerance of different 658 staphylococci in LB containing generated polysulfides from 1 mM NaSH. Plo tted are the mean values of at least 659 three independent, biological replicates. Points show the data of the individual replicates . Squares denote the 660 presence/absence of cst genes following the color scheme from Fig. 1. C) Comparison of the polysulfide tolerance 661 of the analyzed staphylococci from B ) sorted by the presence (+ cst) or absence ( - cst) of cst1. Depicted are box 662 plots without outliers showing the median and the first and third quartile. Whiskers mark the minimum and maximum 663 of the data range. Asterisks indicate statistical significance (****P ≤ 0.0001) from two-tailed Student’s t-tests. Plots 664 in b) and c) follow the general color scheme from Fig. 1 for different Staphylococcus species. 665 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 28 666 Figure 5: Impact of polysulfides on different S. aureus strains.A) Heat maps displaying the relative polysulfide 667 tolerance of multiple S. aureus wild-type strains at different polysulfide concentrations (“low”, generated from 1 mM 668 NaSH; “high”, generated from 3 mM NaSH ) and chain lengths ( 1 mM of Na2S2, Na 2S3, or Na2S4) after 16 h of 669 incubation. Depicted are the mean values of at least three independent, biological replicates. Squares and circles 670 indicate the presence/absence of cst1 or cst2 genes, following the color scheme from Fig. 1. MRSA strains are 671 marked with asterisks in front of their strain designation. B) Relative polysulfide tolerance of the tested strains from 672 A) grouped by harboring cst1 + cst2 (n=6), only cst1 (n=9), or none (n=2). The box plots show the median and the 673 first and third quartiles. Whiskers mark the minimum and maximum of the data range. Asterisks indicate statistical 674 significance (ns = not significant; *P = 0.05 to 0.01; ** P = 0.01 to 0.001; *** P = 0.001 to 0.0001; **** P ≤ 0.0001) 675 from two-tailed Student’s t-tests. C) Heat maps displaying the relative polysulfide tolerance of S. aureus cst-mutant 676 strains and their respective wild types at different polysulfide concentrations ( “low”, generated from 1 mM NaSH; 677 “high”, generated from 3 mM NaSH) and chain lengths (1 mM of Na2S2, Na2S3, or Na2S4) after 16 h of incubation. 678 Plotted are the mean values of at least three independent, biological replicates. Squares and circles indicate the 679 presence/absence of cst1 or cst2 genes, following the color scheme from Fig. 1. MRSA strains are marked with 680 asterisks in front of their strain designation. D) Relative polysulfide tolerance of the tested strains from C) grouped 681 by their cst configuration (cst1 + cst2 n=3, cst1 n=3, and ∆cst1 n=2) and depicted as box plots. Shown are the 682 median and first and third quartiles. Whiskers mark the minimum and maximum of the data range. Statistical 683 significance from unpaired two-tailed Student’s t-tests is denoted as asterisks ( ns = not significant; *P = 0.05 to 684 0.01; **P = 0.01 to 0.001; ***P = 0.001 to 0.0001). 685 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint 29 686 Figure 6: Competitive growth behavior of MRSA (COL and N315) and MSSA (Newman) against the MSSA 687 strain RN4220 during serial passaging over the course of 72 h in medium with or without 1 mM Na 2S3. 688 Depicted are the mean values of three independent, biological replicates. Error bars indicate SD. Statistical 689 significance from unpaired two-tailed Student’s t-tests is denoted as asterisks ( ns = not significant; *P = 0.05 to 690 0.01; **P = 0.01 to 0.001; ***P = 0.001 to 0.0001; ****P ≤ 0.0001). 691 692 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 28, 2026. ; https://doi.org/10.64898/2026.01.28.702267doi: bioRxiv preprint